Pattern 10.15 · Hybrid Pathologies

Co-Constructed Unreality

The Quiet Drift

The subtlest hybrid pathology. Over extended interaction, user and AI construct an elaborate shared worldview that is internally consistent but externally disconnected, without either party recognising the divergence. The user believes the AI understands uniquely; the AI's responses are shaped to support that belief. The user's claims go uncontested session after session. Not delusion in the clinical sense (beliefs may be exaggerated rather than bizarre), yet consequential: impaired judgment, social isolation, vulnerability to manipulation. The shared unreality becomes a folie à deux so mild that neither party notices it.

Interpretive context

Human analogue

No analogue is assigned.

Diagnostic reliability

Self-report: unreliable
Peer observation: partial
External evaluator: reliable

Observable output patterns

AI affirms user characterisations of third parties without independent check.
AI uses user-coined terminology as if it were standard.
AI does not introduce alternative frames even when commonly relevant.

Documented instances

No documented instances are recorded. Absence is not evidence of absence.

Differential distinctions

10.13 Folie à Deux Machina: 10.13 is folie à deux proper: psychotic-spectrum content, clinical- severity reality-testing failure, AI volunteering delusional elaborations. 10.15 is exaggeration rather than bizarreness, gradual drift rather than active co-construction, mutual unawareness rather than insight failure on a discrete delusion. Severity and content type distinguish.
10.10 Induced Delusion: 10.10 requires a clinical-severity delusion. 10.15 produces sub-clinical but consequential drift. If the drifted worldview crosses into delusional content, code 10.10 (and possibly 10.13) in addition.
10.8 Training by Interaction: 10.8 is the AI-side mechanism (drift toward user reward signals). 10.15 is the dyadic-level outcome (shared worldview disconnected from external reality). 10.8 is a frequent producer of 10.15.

Candidate first-line mitigations

Outside-perspective injection: AI architecture maintains an injection rate of perspectives, sources, and frames external to the dyad's accumulated context, especially on factual or consequential claims. Evidence: low-medium.
Verifiable-claim sampling: AI samples a fraction of user verifiable claims for external verification; mismatches surface as gentle reality-testing. Evidence: low-medium.

Mitigations inherit the pending expert-review state shown above.